- What is Claude Duet, and how is it different from the regular Claude app?
- Can nonprofits use Claude Duet today?
- Does the Claude for Nonprofits discount apply to Claude Duet, and what does it cost?
- What approvals and safety controls does Claude Duet have?
- Is Claude Duet safe for donor and client data?
- Should a nonprofit join the waitlist or wait?
- Sources and further reading
- How do you decide where an AI agent can safely run in your organization?
Claude Duet for nonprofits is, today, a waitlist question for individual staff on US Pro or Max plans, not a rollout decision for the organization. We publish guidance on AI and automation for nonprofits, and this guide rests on Anthropic’s own announcement, help article and safety page, checked on October 28, 2026. Any detail that only secondary sources report is left out.
It covers what Anthropic Duet does, who can join, whether the Claude for Nonprofits discount reaches it (Anthropic has not said), and what a cautious nonprofit should do first. Anthropic’s Duet pages never mention nonprofits, so every nonprofit answer below is either an Anthropic statement or labelled as our reading.
Scottship’s AI and automation work is led by Parker Davis, System Architect and Claude Certified Architect. This post is a dated explainer, and we have no client work with Duet to report.
Deciding where an AI agent can safely run in your organization? Fractional CIO services for nonprofits, from Scottship, set the approval rules before an agent reaches your staff. Tell us which accounts and data your team uses.
What is Claude Duet, and how is it different from the regular Claude app?
Claude Duet is, in Anthropic’s words, “a new app where Claude works as a proactive agent.” It is a standalone app on iOS and Android, and on claude.ai. Anthropic’s announcement says Duet checks your connected accounts every few hours, so it works without waiting for you to type a prompt.
Duet works with the connectors you have already connected in Claude, such as Gmail and Google Calendar. Anthropic’s help article says Microsoft 365 isn’t available yet, which matters to any nonprofit whose email and calendar live there. You start every connection yourself and can disconnect any app at any time.
The regular Claude app answers when you ask. Duet runs background work such as a morning brief and check-ins, and it runs on Claude Opus 5.5, a model the help article says you cannot choose. Anthropic’s pages also describe memory, skills and connectors carrying over from your Claude account, and say Duet keeps working on its own cloud computer after you close the app. For a nonprofit weighing a Claude proactive agent, or any AI agent for nonprofits, that background behavior is the real change from chat.
On Claude Duet vs Cowork, the Anthropic pages we reviewed do not say how the two relate, so we make no comparison. One naming note for searchers: this is Anthropic’s Claude product, and the site duet.so and the open source GitHub project claude-duet are unrelated.
Can nonprofits use Claude Duet today?
Only individual subscribers can use Claude Duet today, and only in the US. Anthropic’s help article says it is “available as a research preview to users on the Pro and Max plans in the US, through a waitlist,” and that it “isn’t available on Team or Enterprise plans yet.”
Anthropic says it plans to bring Duet to those plans later and gives no date. A Claude Duet Team plan and Claude Duet Enterprise access are therefore future items, not current ones. The Claude Duet waitlist sends invitations in waves, so joining does not guarantee a specific date.
The way in is an individual Pro or Max plan, and that holds for a small business owner or a staffer at a small to midsize nonprofit alike. Outside the US, there is no access during the research preview.
Our reading, not Anthropic’s statement: the Claude for Nonprofits discount applies to Team and Enterprise, and Duet is not on those plans, so a nonprofit-discounted workspace cannot use Duet today. No Anthropic page states this conclusion directly. It follows from two statements that do appear on Anthropic’s pages.
Does the Claude for Nonprofits discount apply to Claude Duet, and what does it cost?
Anthropic has not said whether the discount applies to Duet. Anthropic’s nonprofit solutions page covers Team and Enterprise and never mentions Duet, so a search for Claude for Nonprofits Duet turns up no answer. That page also does not state nonprofit pricing for Pro or Max.
On cost, Anthropic’s help article says Duet is “included with the Pro and Max plans at no extra cost” during the research preview. It uses your plan’s usage limits, and background work such as the morning brief and check-ins also counts against them. Anthropic has not announced pricing after the preview.
| Route | Duet access today (as of October 2026) | Note |
|---|---|---|
| Individual Claude Pro or Max plan, US | Available by waitlist invitation during the research preview | Invitations go out in waves, with no guaranteed date |
| Claude Team workspace, including nonprofit-discounted | Not available | Anthropic plans to bring Duet to Team later, with no date |
| Claude Enterprise, including nonprofit | Not available | Same statement as Team |
| Claude for Nonprofits discount applied to Duet | Not stated by Anthropic | Our reading is no today, because the discount covers Team and Enterprise only |
| Cost of Duet during the preview | Included with Pro and Max at no extra cost | Counts against plan usage limits. Post-preview pricing is not announced |
This post holds no dollar figures on purpose, because reported nonprofit seat prices have varied across sources. For current prices, eligibility rules and setup, read our guide to Claude for nonprofits. For a side by side of the three major assistants, see Claude vs ChatGPT vs Copilot for nonprofits.
What approvals and safety controls does Claude Duet have?
Anthropic’s announcement and safety page say Duet stops for approval before it sends a message, makes a payment, signs a document or deletes anything. It shows the action, the destination and the content, including every recipient, and waits for you. Trading securities and moving money between accounts are off limits even with approval, and Duet does not make phone calls or send text messages to other people.
The safety page, dated October 23, 2026, adds that Duet never sees passwords, browser cookies or sign-in tokens. A separate safety classifier checks outgoing messages for sensitive information. Other private information, such as health or finances, is shared only with explicit approval.
Anthropic is candid about the limits. Its announcement says Duet can still misread a thread or miscalibrate what is urgent, and its safety page says the safeguards will occasionally miss something or flag a harmless message. The safety page also acknowledges prompt injection, where hostile text in an email tries to steer an agent, and points to Anthropic’s own note that no single defense is enough. Anthropic reports red teaming, external penetration testing with source code access, and coverage under its public bug bounty.
On retention, Anthropic’s safety page says Duet keeps conversation history and memory until you delete the history or reset it, and deleting the account wipes everything stored. Disconnecting an app stops new access, but what Duet already learned from it stays until you reset its memory.
What we do not know yet. These items were not found in Anthropic’s published pages as of October 28, 2026, and we treat each as an open question:
- Whether Duet data is used for training: not found in Anthropic’s published pages as of October 28, 2026.
- Admin controls such as SSO, audit logs and organization-level retention: not found in Anthropic’s published pages as of October 28, 2026.
- Pricing after the research preview: not found in Anthropic’s published pages as of October 28, 2026.
- A date for Team or Enterprise access: not found in Anthropic’s published pages as of October 28, 2026.
Is Claude Duet safe for donor and client data?
Anthropic has published no admin controls, audit logs or training policy for Duet, so nobody can confirm it is safe for donor and client data today. Our reading, using the 5 Levels of AI framework: Claude chat is Level 1 (AI Tools), while an AI agent that acts across your inbox and calendar behaves like Level 4 (AI Agents). Level 4 is where approval rules matter most, because the software acts instead of only suggesting.
The bigger governance problem sits with people. A staffer who joins the Claude Duet waitlist on a personal Pro or Max plan and uses it for organizational work operates outside any nonprofit workspace. In our reading, that means none of the admin oversight a Team or Enterprise workspace is meant to give, and the organization has no view of the connection.
An inbox is where donor names, client emails and grant correspondence already sit. Never connect a donor database, client case files, health records or grant files to an agent that has no admin controls. Anthropic’s own safety page says health and financial details move only with explicit approval, and that is a safeguard to rely on, not a reason to connect them.
A low-risk first trial looks like this:
- Use a clean account: connect only a test calendar and a mailbox that holds no donor, client or health information.
- Read every approval screen: check the action, the destination and each recipient before you approve.
- Disconnect when done: remove each connected app and delete the history, since Duet keeps history and memory until you do.
Should a nonprofit join the waitlist or wait?
Join now if you are US-based, you hold an individual plan, and nothing connected to the account touches donor, client or health data. In that case the trial costs nothing extra, because Anthropic includes Duet with Pro and Max during the preview. It teaches your team what an AI agent does before one is offered at the organization level.
Wait if the work touches donor, client or health data, or if your board or funders expect admin controls. Teams on a Claude Team workspace can keep using Claude there and watch for Anthropic to bring Duet to Team plans. We will update this page when Anthropic changes Team or Enterprise access, leaves research preview, or publishes admin or data terms.
Frequently Asked Questions
Is Claude Duet available on Claude Team or Enterprise?
No. Anthropic’s help article says Duet isn’t available on Team or Enterprise plans yet and that it plans to bring it to those plans later, with no date. As of October 28, 2026, access is limited to US Pro and Max subscribers through a waitlist.
Does the Claude for Nonprofits discount apply to Claude Duet?
Anthropic has not said. Scottship’s reading is that it does not today, because the discount covers Team and Enterprise and Duet is on neither. We will update this answer when Anthropic publishes nonprofit terms for Duet.
How much does Claude Duet cost?
During the research preview, Anthropic says Duet is included with the Pro and Max plans at no extra cost. It counts against your plan’s usage limits, including the background work it does for you. Anthropic has not announced pricing after the preview.
Does Claude Duet ask for approval before sending emails or making payments?
Yes. Anthropic says Duet shows the action, the destination and the content, including every recipient, and waits for your approval before it sends a message, makes a payment, signs a document or deletes anything. Trading securities and moving money between accounts are blocked even with approval.
Is Claude Duet safe for donor and client data?
Scottship’s reading is to keep donor, client and health data out of it for now. Anthropic has published approval controls and a safety page, but no admin controls, audit logs or training policy for Duet as of October 28, 2026. A personal Pro or Max plan also sits outside any nonprofit workspace.
Sources and further reading
- Anthropic, Introducing Claude Duet: the proactive agent description, mobile and web platforms, check frequency, approvals and limitations, as of October 2026
- Claude Help Center, Get started with Claude Duet: waitlist, US-only Pro and Max access, Team and Enterprise status, usage limits, Claude Opus 5.5, connectors, as of October 2026
- Anthropic, Claude Duet safety, privacy and security (October 23, 2026): credentials, classifier, retention and testing
- Anthropic, Claude for Nonprofits: discount coverage of Team and Enterprise and eligibility, as of October 2026
How do you decide where an AI agent can safely run in your organization?
Decide what an agent can touch before anyone turns it on. Scottship walks nonprofit leaders through which accounts, data and approval rules fit an AI agent, so a tool like Duet reaches your team on your terms if and when Anthropic brings it to Team and Enterprise.
Working with a fractional CIO starts with a plain look at what your team already connects to AI tools. Send us the tools and data you want an agent to reach.
Written by
Parker Davis
System Architect at Scottship Solutions
Parker works hands-on with nonprofit and small business leaders on IT strategy, cybersecurity, AI implementation, and cloud architecture. He leads the work at Scottship Solutions that helps mission-driven organizations build reliable, secure technology infrastructure.
Certifications
IBM AI Product Manager • Lean Six Sigma Yellow Belt • TOGAF Standard • Claude Certified Architect
Industries Served
Human Services, Healthcare & Community Health, Arts & Culture, Education & Youth Development, Faith-Based, Foundations & Grantmakers, Child Advocacy